The bayesian network analysis model of the diversity culture path of the garden talents

The scientific and efficient cultivation of gardening professionals is an important direction of the current education and teaching work carried out in higher vocational colleges and universities [1]. As far as higher vocational colleges and universities are concerned, the cultivation of gardening professional talents focuses on the acquisition and enhancement of practical operation skills, which is also one of the characteristics of the cultivation of talents in higher vocational colleges and universities [2]. Therefore, in the cultivation of gardening professionals in higher vocational colleges and universities, it is of great practical significance to promote the close combination of theory and practical teaching, and to carry out practical teaching with theory as the guide [3-4].

However, in the actual teaching of gardening majors in higher vocational colleges and universities, there is a disconnect between theory and practice teaching. First of all, some teachers in the gardening professional teaching, is still limited to the traditional teaching concepts of bondage, teaching lack of innovation, theory teaching is still overriding the practical teaching, resulting in students’ practical training ability training failed to meet expectations [5-7]. Secondly, some teachers are actively promoting the combination of theoretical and practical teaching, but there is a phenomenon of two skins, the students’ practical training operation not only can not consolidate the knowledge they have learned, but also produce blindness to the students’ later learning [8-10]. Furthermore, higher vocational colleges and universities in the gardening professional practical training operation, there are some schools do not pay enough attention to the school-enterprise cooperation mechanism is not perfect and other problems, resulting in the gardening talent diversified training mechanism is not sound [11-13]. In view of the above problems, information technology is integrated into the composition of diversified training of gardening talents, improve the school-enterprise cooperation mechanism, broaden the scope of cooperation, so as to provide students with long-term, stable practical training operation opportunities [14-16]. Based on the current construction problems, put forward targeted optimization countermeasures, so that the innovative construction of talent cultivation path is put into practice.

Research and construction of the curriculum system reconstruction, teaching mode innovation and industry-teaching integration of practical training as the main “teaching” path system to cultivate gardening talent skills. The path system of “educating”, which is mainly based on the shaping of environmental education, the nourishment of ideological and political elements and the leadership of party members, cultivates the quality of gardening talents. In order to verify the satisfaction of the “education and education integration” path, a comprehensive evaluation index system for the diversified cultivation path of gardening talents was constructed. The effectiveness of the model was verified in terms of accuracy, precision, recall, specificity, and false-positive rate. A comprehensive evaluation model based on Bayesian network was constructed to analyze the satisfaction level of the pathway of “education and education integration”. Based on the analysis results, the constructed pathway is further optimized.

2

The diversified cultivation path of gardening talents based on the “integration of education and training”

2.1

The “teaching” pathway system

Focusing on “teaching”, we construct a path system mainly consisting of the path of curriculum reconstruction, the path of teaching mode innovation and the path of industry-teaching fusion and practical training, so as to build up a platform and find the right hand for efficient “teaching”. 1)

Implementing the path of curriculum system reconstruction, optimizing the curriculum framework of garden craftsmen.

According to the law of career growth and job cognition of garden craftsmen, the school and enterprise jointly build a “double subject, three systems, and four advanced” curriculum system with distinctive occupational characteristics and highlighting the work process orientation, which respectively carries the logical progressive relationship of “from basic to professional”, “from simple to complex”, “from recognition to practice”, and the professional connotation of different stages of “foundation-professional-extension-expansion”, “interest-single-comprehensive-innovation”, “perception-absorption-training-competency-creation”. The different contents alternate between schools and companies.

2)

Implementation of teaching mode innovation path and activation of teaching methods for garden artisans

Implement the work-oriented teaching mode with distinctive characteristics of the gardening professional group, and lead the reform of teaching methods for the cultivation of garden craftsmen. According to the law of professional growth and the law of cognition of jobs in the gardening industry, combining with the curriculum, projects and professional practice, the cultivation of craftsmen’s ability is carried out through the four progressive ability cultivation lines of “works at the curriculum level - works at the cognitive level - works at the project level - works at the professional level”, and through the process of formation of a number of works at the professional level, we can construct a teaching method suitable for the cultivation of high quality gardening industry jobs. Through the formation process of several professional level works, the teaching mode is constructed to cultivate high-quality garden artisans who are required by the posts in the gardening industry. The implementation of each section is carried out in accordance with the 3-step process of “accepting the work - implementing and completing the work - evaluating the work”, and the assessment and evaluation is standardized in accordance with the 2 evaluation systems of “quality of the work and personal professional qualities”.

3)

Implementing the practical training path of industry-teaching integration, and realizing the engineering-learning combination of gardening craftsmen.

Relying on flower factories, landscape design centers, “project-based” training bases, and school-enterprise garden industry colleges, the company has built a “five-star, project-based, industry-teaching” training platform to carry out the entrepreneurial operation of practical training projects. Introducing enterprise management concepts, constructing a new mode of implementing the integration of production and education of garden artisans and a new evaluation mechanism, leading garden artisans to undertake tasks as employees of enterprises, and building a platform for practicing the ability of integration of production and education of garden artisans. Schools and enterprises jointly build industrial colleges, promote intangible cultural heritage into the campus, inherit the classic skills, and help industrial development [17].

2.2

“Education” pathway system

Focusing on “education”, focusing on “cultivating people + morality”, we will build a path system with the path of environmental education shaping, the path of ideological and political elements and the path of party member vanguard leadership as the main body, so as to clarify the carrier and find the right direction for efficient “education”. 1)

Implement the path of environmental education and create a professional atmosphere for garden craftsmen

Create an environmental education space with “school-enterprise integration, integration of production and education”, “clear etiquette and righteousness, fine skills, and morality” as the main content, and match the characteristics of the garden profession, give full play to the effectiveness of environmental education, and lead the direction of environmental education.

2)

Implement the nourishing path of ideological and political elements to guide the value orientation of garden craftsmen

Create a brand of ideological work of “Ideological and Political Theory”, carry out a series of hot topics of “Ideological and Political Thinking” according to the characteristics of different stages of students’ growth, and promote them through the WeChat public account of the University Counselor Alliance, so as to resonate with young students in a new form of ideological work, arouse students’ thinking, and guide students’ growth. Use the waiting time to implement the ideological and political section before class of “talking about current affairs 5 minutes before class”. From the subtleties, the ideological and political concepts of the course and the core values of socialism are integrated into the whole process of the training of garden craftsmen, so as to play the educational effect of “moisturizing things silently”.

3)

Implement the pioneer leading path of party members and highlight the exemplary leading role of garden craftsmen

In line with the concept of “education changes destiny and employment is related to life”, we will build a platform for the employment and entrepreneurship ability of college students in the landscape professional group, excavate the advanced nature of professional student party members in employment and entrepreneurship skills from the perspective of entrepreneurship and employment, and find out the integration point of college student party member training and skilled talent training from the perspective of multiple education, so that students can become garden craftsmen who can “do things, know how to do things, and do good things” in the future, and truly form a strong driving force for all students in terms of employment and entrepreneurship.

3

Bayesian network based path evaluation

3.1

Bayesian networks

3.1.1

Overview of the model

A Bayesian network is a directed acyclic graph (DAG) consisting of nodes and directed edges. The nodes in the network represent random variables and the directed edges represent probabilistic dependencies between the variables. Each node is accompanied by a conditional probability table (CPT), which quantitatively presents the probabilistic dependencies between the variables [18].

Bayesian networks are often denoted as B = 〈G(X, A), Θ〉, G represents a directed acyclic graph, $X = {X_{1}, X_{2}, \dots, X_{n}}$ represents a set of random variables, A is a directed edge, and $Θ = {θ_{X_{1}} | U_{1}, θ_{X_{2} | U_{2}}, \dots, θ_{X_{n} | U_{n}}}$ is a network parameter. Where U_i is the set of parent nodes of X_i and $θ_{x_{i}} | U_{i}$ is the conditional probability table corresponding to it. Knowing the network parameters, the joint probability of all nodes can be calculated according to Bayes’ theorem, i.e.: (1) $P (X_{1}, X_{2}, \dots, X_{n}) = \prod_{i = 1}^{n} P (X_{i} | U_{i}) = \prod_{i = 1}^{n} θ_{X_{i} U_{i}}$

In practice, the learning of Bayesian networks mainly includes structure learning, parameter learning and network inference.

3.1.2

Bayesian network structure learning

Network structure learning is a process of obtaining network structure from a dataset D. The commonly used construction methods are categorized into expert experience method and machine learning method. Due to the large number of variables involved in the study, in this paper, machine learning method will be used to automatically obtain the network structure of bus service under the effect of multiple variables. The commonly used structure learning algorithms are divided into 2 categories: 1)

Search scoring based algorithms, i.e., searching for the optimal network structure using scoring functions.

2)

Constraint-based algorithms, also known as algorithms based on dependency analysis, which realize structure learning by testing the dependency relationship between nodes through conditional independence (CI).

The learning process of search scoring algorithm is relatively simple, the search space is large, and the learning efficiency is relatively high, in this paper, we will use the search scoring algorithm GTT to learn the network structure. The algorithm uses the scoring function to measure the fit of structure G and data set D to seek the optimal network structure. The specific steps are as follows: 1)

Network plus edge. The network is initially a blank structure, and searches for and adds directed edges that are beneficial to the network structure scores until the structure scores are no longer increasing.

2)

Network minus edges. On the basis of the network structure in the previous step, the directed edges that have no positive effect on the network structure are continuously searched and removed until the structure score no longer increases. The network structure merit function uses a K₂-score function, i.e.: (2) $K_{2} (G, D) = \sum_{i = 1}^{n} \sum_{j = 1}^{q_{i}} (\log (\frac{(r_{i} - 1)!}{(N_{i j} + r_{i} - 1)!}) + \sum_{k = 1}^{r_{i}} \log (N_{i j k}!)) + \log (P (G))$

where P(G) is the a priori probability of the network structure. r_i is the number of values taken by node X_i. q_i is the combination of values taken by the parent node of node X_i. N_ij is the number of node X_i parent nodes in the data set that are in the kth value combination and $N_{i j} = \sum_{k = 1}^{r_{i}} N_{i j k}$ .

Compared with general search scoring algorithms, K₂ the algorithm has the following advantages: 1)

The algorithm does not need to give the order between network nodes in advance during structure learning.

2)

The removal of redundant edges in the algorithm reduces the occurrence of model overfitting situations.

3)

The algorithm has high learning efficiency and accuracy.

3.1.3

Bayesian network parameter learning

Based on the constructed Bayesian network structure G and data D, the parameters of the Bayesian network are learned to obtain the conditional probability distributions of the node variables $θ_{x_{i} | U_{i}}$ In this paper, the most commonly used Maximum Likelihood Estimation (MLE) method is used to compare the likelihood of the samples and the parameters to learn the model network parameters.

3.1.4

Bayesian network inference

Inference in Bayesian networks is implemented on the basis of an established network structure and a conditional probability distribution table to calculate the probability of an event occurring. The applied conditional probability theory formula is: (3) $P (A | B) = \frac{P (A, B)}{P (B)}$

where P(A|B) is the probability of A occurring under the given B conditions. P(A, B) is the joint probability distribution of A and B. P(B) is the marginal probability of B.

In order to facilitate effective reasoning, this paper adopts the confidence propagation algorithm to calculate the probability distribution P(X, e) of each variable in the network after the input of new evidence e The confidence propagation algorithm belongs to a kind of message propagation algorithm, which mainly relies on probabilistic dependencies between nodes to realize the transmission of information, and the specific formula is as follows: (4) $P (X U, e) = λ_{e} (X) \prod_{i} π_{x} (U_{i}) \prod_{j} λ_{r_{j}} (X)$ (5) $P (U_{i}, e) = \sum_{X U, {U_{i}}} λ_{°} (X) θ_{X U} \prod_{k \neq i} π_{x} (U_{k}) \prod_{j} λ_{Y_{j}} (X)$ (6) $P (X, e) = \sum_{U} λ_{e} (X) θ_{X_{U}} \prod_{i} π_{x} (U_{k}) \prod_{k \neq j} λ_{Y_{k}} (X)$

where U is the set of parent nodes of X. Y is the set of child nodes of X. π_x(U) represents the information passed from node U to X. λ_Y(X) is the information passed from X to Y. λ_e(X) is the evidence input indicator, which is 1 if e is entered in the X variable and 0 otherwise. P(XU, e) is the joint probability distribution after inputting evidence. P(U, e) is the marginal probability of U_i after input e. P(X, e) is the marginal probability of X.

3.2

Evaluation index system for diversified cultivation path of gardening talents

The evaluation index system of the diversified cultivation path of landscape talents based on “education and teaching integration” constructed in the study is shown in Table 1. The first-level indicators include “teaching” path and “education” path. There are 6 secondary indicators and 18 tertiary indicators under the primary indicators.

Table 1.

The landscape talent diversity culture path evaluation index system

Primary index	Secondary index	Tertiary index
Teaching(A)	Curriculum reconstruction(A1)	The degree of docking of the course and the industry(A11)
		The rationality of the curriculum structure(A12)
		The richness of the course resources(A13)
	Teaching model innovation(A2)	Diversity of teaching methods(A21)
		The informationization of teaching technology(A22)
		The participation of teaching activities(A23)
	Training and integration of production(A3)	The quantity and quality of cooperative enterprises(A31)
		Practice and innovation of training programs(A32)
		Students’ satisfaction with training(A33)
Instruction(B)	Environmental shaping(B1)	Optimization of the environment(B11)
		The frequency of students’ participation in cultural activities(B12)
		Environmental effect(B13)
	The elements of thinking are moist(B2)	The coverage of the government course(B21)
		The harmony of the education and the professional education(B22)
		Students’ improvement of political literacy(B23)
	Leader of party members(B3)	Demonstration of party members pioneer(B31)
		The effect of teaching and education(B32)
		Job creation(B33)

3.3

Comprehensive evaluation model based on Bayesian network

The Bayesian network model that combines probability theory and graph theory is a data mining method that can explore the complex relationship between variables. Therefore, with the help of Bayesian network, this paper constructs the joint mechanism of the subjective and objective levels of the index system for the evaluation of diversified cultivation paths of gardening talents. Among them, the evaluation results of subjective indexes are easily disturbed by other factors and are latent variables that cannot be directly measured. While the evaluation results of objective indicators as explicit variables can be obtained through direct measurement and scoring by experts. Therefore, the objective evaluation which is easy to obtain is regarded as the a priori distribution, based on which the subjective evaluation and objective evaluation can be combined and constructed into the posterior distribution with the help of Bayes’ theorem, and the average value of the expectation of the posterior distribution is the comprehensive score. The specific steps of the comprehensive evaluation model based on Bayesian network are as follows: 1)

Determination of index weights

Determine the weights of the indicators in the comprehensive evaluation index system based on the Delphi method w_i [19].

2)

Quantification of indicators based on scales

For objective indicators, data are collected through on-site research and scored directly by professionals in the field based on different industry standards and their own experience, so numerical scales are used. For subjective indicators, it is necessary to issue questionnaires to the gardening talents to consult their feelings about the indicators, on the basis of which qualitative descriptions are assigned corresponding scores, so the Likert scale is used. The set of evaluation options of Likert scale is {very dissatisfied, dissatisfied, average, satisfied, very satisfied}, and the set of corresponding scores is {2, 4, 6, 8, 10}.

3)

Bayesian network construction

Established research defaults to the indicators are independent of each other. The actual subjective evaluation of any indicator of gardening talent scores are the result of the interaction between the actual implementation efficiency of each indicator and the joint impact, the plain Bayesian network model is more idealized, not taking into account this interaction relationship, which will lead to bias in the subjective scoring, and therefore it is not appropriate to use the plain Bayesian network for subjective evaluation. The correlation between all objective indicators, and the three-level subjective indicators under different second-level indicators is small, and can be regarded as independent of each other. The correlation between the three-level subjective indicators under the same second-level indicators is large, and the interaction relationship between the indicators cannot be ignored. Therefore, the subjective evaluation sub-Bayesian network is constructed according to the semiplain Bayesian network, the objective evaluation sub-Bayesian network is constructed according to the plain Bayesian network, and the two sub-Bayesian networks are combined to form a comprehensive evaluation Bayesian network.

4)

Bayesian network inference

The joint probability distribution of the three-level indicators can be obtained according to the statistical results of the numerical scale and the Likert scale, and the marginal probability distribution of each three-level indicator is calculated on this basis: (7) $P (X_{i}^{I I I}) = \sum_{X_{A 13}^{I I I}} \sum_{X_{A 17}^{I I I}} \dots \sum_{X_{m}^{I I I}} \dots \sum_{X_{B 41}^{I I I}} \sum_{X_{B 42}^{I I I}} P (X_{A 13}^{I I I}, X_{A 17}^{I I I} \dots X_{m}^{I I I} \dots, X_{B 41}^{I I I}, X_{B 42}^{I I I}), m \neq i$ (8) $P (Y_{i}^{I I I}) = \sum_{Y_{A 11}^{I I I}} \sum_{Y_{A 12}^{I I I}} \dots \sum_{Y_{m}^{I I I}} \dots \sum_{Y_{B 11}^{I I I}} \sum_{Y_{B C}^{I I I}} P (Y_{A 11}^{I I I}, Y_{A 12}^{I I I} \dots Y_{m}^{I I I} \dots, Y_{B 41}^{I I I}, Y_{B 42}^{I I I}), m \neq i$ (9) $P (X_{i}^{I I I} | Y) = \sum_{X_{A 13}^{I I I}} \sum_{X_{A 17}^{I I I}} \dots \sum_{X_{w}^{I I I}} \dots \sum_{X_{B 4}^{I I I}} \sum_{X_{B 42}^{I I I}} P (X_{A 13}^{I I I}, X_{A 17}^{I I I} \dots X_{m}^{III} \dots, X_{B 41}^{III}, X_{B 42}^{III} | Y), m \neq i$

Where $X_{i}^{I I I}$ is the rating of the three-level subjective indicator i, $X_{i}^{I I I} \in {2, 4, 6, 8, 10}$ , $Y_{i}^{I I I}$ is the rating of the three-level objective indicator i, $Y_{i}^{I I I} \in {2, 4, 6, 8, 10}$ , Y is the objective rating.

Using Bayesian inference to calculate the marginal probability distribution of second-level indicators, first-level indicators, subjective and objective ratings in turn according to the rules of calculation between the indicators at each level, and the rules of calculation of inference between the indicators at each level of the subjective evaluation sub-Bayesian network are as follows: (10) $P (X_{j}^{I I I}) = P (\frac{\sum ω_{i}^{I I I} X_{i}^{I I I}}{\sum ω_{i}^{I I I}})$ (11) $P (X_{k}^{I}) = P (\frac{\sum ω_{j}^{I I} X_{j}^{I I}}{\sum ω_{j}^{I I}})$ (12) $P (X) = P (\frac{ω_{A}^{I} X_{A}^{I} + ω_{B}^{I} X_{B}^{I}}{ω_{A}^{I} + ω_{B}^{I}})$

In the formula, the third-level subjective indicator i is subordinate to the second-level subjective indicator j, and the second-level subjective indicator j is subordinate to the first-level subjective indicator k. X is the subjective rating, $X_{k}^{I}, ω_{k}^{I}$ is the rating and weight of the first-level subjective indicator k, $X_{j}^{I I}, ω_{j}^{I I}$ is the rating and weight of the second-level subjective indicator j, and $ω_{i}^{Π}$ is the weight of the third-level subjective indicator i.

Objective evaluation sub-Bayesian network, subjective evaluation sub-Bayesian network under the condition of known objective ratings between the levels of indicators inference calculation rules are the same as this. 5)

Calculation of posterior distribution

The use of Bayes’ theorem to combine subjective evaluation, objective evaluation, and get the objective evaluation of the posterior probability distribution function: (13) $P (Y | X) = \frac{P (X, Y)}{P (X)} = \frac{P (Y) P (Y | X)}{P (X)}$

6)

Composite score

The posterior probability distribution function is known to solve for the expectation of the posterior distribution E(Y|X = x): (14) $E (Y | X = x) = \sum_{y} y P (Y = y | X = x) = \sum_{y} y \frac{P (Y = y) P (X = x | Y = y)}{P (X = x)}$

E(Y|X = x) is a discontinuous function about X, then the composite score χ is: (15) $χ = E (Y | X = x) = \sum_{x} \sum_{y} y \frac{P (Y = y) P (X = x | Y = y)}{P (X = x)} / n$

Where, n is the number of values taken by the subjective score X.

The evaluation set of diversified cultivation paths for gardening talents is defined as {poor, fair, good, excellent}, and the corresponding set of comprehensive rating thresholds is {[5, 6), [6, 7), [7, 8), [8, 9)}.

4

Model validation and satisfaction analysis

4.1

Validation of comprehensive evaluation model based on Bayesian network

4.1.1

Indicators for evaluating the effectiveness of the model

The confusion matrix can usually be used to evaluate the classification effect of Bayesian network models. For a common binary classification situation (e.g., “Category I” and “Category II”), the confusion matrix shown in Table 2 can be obtained.

Table 2.

Confusion matrix

True value	Predictive value
True value	Categories I	Categories II
Categories I	TP	FN
Categories II	FP	TN

Based on the above confusion matrix, the following comprehensive quantitative metrics can be further obtained to assess the classification effectiveness of the model. 1)

Accuracy. Percentage of predicted “Category I” and “Category II” and actually observed “Category I” and “Category II”. Percentage of total number of observations: (16) $A c c u r a c y = \frac{T P + T N}{T P + T N + F P + F N}$

2)

Precision rate. The proportion of predicted “category I” and actually observed as “category I” as opposed to predicted “category I”: (17) $\Pr e c i s i o n = \frac{T P}{T P + F P}$

3)

Recall. Reflects the proportion of predicted “category I” that are actually observed as “category I” as opposed to actually being “category I”: (18) $Re c a l l = \frac{T P}{T P + F N}$

4)

Specificity. Reflects the proportion of samples actually observed as “category II” that the model recognizes as also being “category II”: (19) $S p e c i f i c i t y = \frac{T N}{T N + F P}$

5)

False positive rate. Reflects the proportion of observations that are actually “Category II” but are recognized as “Category I” by the model.

6)

F1 value: The F1 value is the mean of precision and recall:

(20)

F_{1} = \frac{2 \cdot \Pr e c i s i o n \cdot Re c a l l}{\Pr e c i s i o n + Re c a l l}

4.1.2

Model validation

In this paper, K-Fold cross-validation is used, with 200 questionnaires as 1 group and 1000 valid questionnaires divided into 5 groups. In cross-validation, 1 group of data is randomly selected as the validation set, and the rest of the groups are used as the training set, and the cycle is repeated 5 times sequentially. The distribution of satisfaction level after cross-validation is shown in Figure 1. The set of evaluation options is {very dissatisfied, dissatisfied, average, satisfied, very satisfied}. From the figure, it can be seen that the diagonal data i.e., the number of copies of real options and accurate predictions, in which there are 31, 422, 166, 34, 41 copies of very satisfied, satisfied, more satisfied, dissatisfied, and very dissatisfied, respectively, which is a total of 694 copies.

According to Figure 1 and equations (16)~(20) can be obtained by the established comprehensive evaluation model based on Bayesian network each evaluation index is shown in Table 3. The comprehensive evaluation model established in this paper can more accurately identify the degree of satisfaction of the diversified cultivation path of gardening talents, and the overall accuracy rate reaches 81.3%.

Table 3.

Satisfaction of Beels network evaluation results

Satisfaction	Accuracy	Precision	Recall	Specificity	False positive rate	F1
Very satisfied	0.813	0.532	0.277	0.964	0.042	0.343
Satisfaction		0.946	0.982	0.516	0.468	0.953
General		0.935	0.495	0.936	0.002	0.633
Discontent		1	1	1	0	1
Very dissatisfied		1	0.916	1	0	0.954

4.2

Analysis of Satisfaction with Diversified Cultivation Paths for Landscape Architecture Talents

Determine the weight of each indicator in the comprehensive evaluation index system based on the Delphi method w_i. The importance of each indicator of the diversified cultivation path of landscape talents is shown in Figure 2. The rationality of course structure (A12), the number and quality of cooperative enterprises (A31), the improvement of students’ ideological and political literacy (B23) and the demonstration role of party pioneers (B31) all have a weight of 0.07, which has the greatest impact on satisfaction.

The probability of satisfaction using the comprehensive evaluation model is shown in Table 4. It can be seen that the probability of the diversified cultivation path of gardening talents being “excellent” is the largest, as high as 0.92, and the probability of the diversified cultivation path of gardening talents being “poor” is 0.06. Compared with “average” and “good”, the value of dissatisfaction is also larger, which indicates that although the gardening talents are generally satisfied with the diversified cultivation path, there are still some dissatisfied areas, and these dissatisfied indicators can be obtained through the node conditional probability table. These dissatisfaction indicators can be obtained by exploring the conditional probability table of the nodes.

Table 4.

The probability of path satisfaction of garden talent diversity culture

Satisfaction	[5,6)	[6,7)	[7,8)	[8,9)
Probability	0.06	0.02	0.02	0.92

The conditional probability of the “teach” path node when the parent node is satisfaction is shown in Table 5. It can be seen that regardless of the value of satisfaction, the probability of the “teach” path node is more than 90% of the value is generally concentrated in the range of [5, 6), which indicates that the current design of the “teach” path of diversified training of gardening personnel has not been able to satisfy the users, and there is still much room for improvement. This indicates that the currently designed “teaching” path for diversified cultivation of landscape talents has not yet satisfied the users, and there is still much room for improvement.

Table 5.

The conditional probability of the “teaching” path node

Satisfaction	The probability of “teaching” path node condition/%
Satisfaction	[5,6)	[6,7)	[7,8)	[8,9)
Worse	96	4	0	0
General	97	3	0	0
Good	92	8	0	0
Excellence	97	3	0	0

The conditional probability of the “education” path when the parent node is the satisfaction and “teaching” path node is shown in Table 6. When the satisfaction is higher, the probability of “education” path nodes is greater, but it can be seen from Table 5 that the current “teaching” path of diversified cultivation of garden talents is not accepted, so the range corresponding to the probability of “education” path nodes still does not reach a high level, and the maximum numerical range is [6,7], which indicates that if the “teaching” path path is improved, the satisfaction of the “education” path nodes will increase accordingly, and even reach a higher level. If you want to improve overall satisfaction, you must first address certain metrics in the “teach” path.

Table 6.

The conditional probability of the “Instruction” path node

Satisfaction	“Teaching”	The probability of “teaching” path node condition/%
Satisfaction	“Teaching”	[5,6)	[6,7)	[7,8)	[8,9)
Worse	Worse	0	97	3	0
General	General	0	41	59	0
Good	Worse	0	50	45	5
Excellence	General	0	0	78	22
Excellence	Good	0	0	0	100

4.3

The strategy to improve the diversified cultivation path of gardening talents

4.3.1

Precise disciplinary positioning

Due to the level of each school, school history, discipline background is different, resulting in the talent training mode of the gardening profession, there are also differences. At present, many institutions in the curriculum reform and teaching practice exploration, should be combined with their own conditions, accurate discipline development orientation, maintain their own professional characteristics, so that the gardening profession can be enriched and developed.

4.3.2

Clarification of guidelines

Gardening is a very comprehensive discipline, talent training should adhere to the principles of art and science mingling, humanities and science and technology penetration, theory and practice are closely integrated, and will be carried through the whole process of talent training. So that students have a solid theoretical foundation, solid practical application ability, and have the ability of various types of garden planning and design and project construction and management. Cultivate compound specialists with a sense of social responsibility, team spirit and innovative thinking, who can be engaged in landscape planning and design, landscape engineering management, professional research and teaching, and adapt to the needs of the market economy and the development of science and technology.

4.3.3

Development of a specialized curriculum

The comprehensive nature of the horticultural discipline determines its diversity and provides a basis for the cultivation of individualized talents. On the basis of accurately grasping the connotation and development trend of the disciplines and following the principle of “broad foundation, exquisite specialty and distinctive personality”, we build a curriculum system for students to choose on their own, which has commonality and can satisfy students’ interests and aspirations, and is conducive to students’ personality development and lifelong learning, so that they can choose relevant disciplines on the basis of completing the curricula of the disciplines and then combining their personalities and interests. On the basis of completing the curriculum of their own disciplines, students can then combine their personalities and interests with the curriculum of related disciplines, so that the knowledge structure has both breadth, thickness and depth.

5

Conclusion

The study analyzes the diversified cultivation path of horticultural talents based on the “integration of education and training” based on the satisfaction evaluation model of Bayesian network. The model in this paper can more accurately identify the satisfaction level of the diversified cultivation path of gardening talents, with an accuracy rate of 81.3%. The model is used to analyze the diversified training path of garden talents based on the design of “integration of education and education”, and it is found that the “teaching” path has not satisfied users, and the probability of the diversified training path of garden talents based on the “integration of education and education” is “excellent” is 92%. In order to enhance the overall satisfaction of the diversified cultivation path of gardening talents, this paper optimizes the three aspects of disciplinary positioning, guiding ideology and professional curriculum system respectively.

Lingua:: Inglese

Frequenza di pubblicazione:: 1 volte all'anno
Argomenti della rivista:: Scienze biologiche, Scienze della vita, altro, Matematica, Matematica applicata, Matematica generale, Fisica, Fisica, altro

Feed RSS della rivista

The bayesian network analysis model of the diversity culture path of the garden talents

Xiaoyan Zhang

Pubblicato online: 26 mar 2025

Ricevuto: 12 nov 2024

Accettato: 14 feb 2025

DOI: https://doi.org/10.2478/amns-2025-0816

Parole chiaveBayesian network, Evaluation index system, Comprehensive evaluation model, Diversified cultivation paths

© 2025 Xiaoyan Zhang, published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Parole chiave
Bayesian network, Evaluation index system, Comprehensive evaluation model, Diversified cultivation paths